Calcite | Level 5

## How can i obtain a relative risk for my multi variable linear regression analysis of my bio-medical?

I am working on a medical project and need help with coding. Here is my problem:

• My dependent variable is continuous ('Y').
• My independent variables include a mixture of about 10 continuous and categorical variables, with 'X' being the main variable of interest.
• My goal is to determine the relative risk for 'X' while controlling for covariates (A, B, C, D).

I tried using `PROC GLM`, but it does not provide relative risk. I understand that `PROC GENMOD` is typically used for categorical dependent variables.

Could someone provide an example of SAS code to calculate the relative risk for my variable 'X'?

2 REPLIES 2
Super User

## Re: How can i obtain a relative risk for my multi variable linear regression analysis of my bio-medi

You could use
proc logistic + /param=glm + lsmean trt/ ilink diff + %nlmeans
Check the second example:
https://support.sas.com/kb/23/003.html
Lapis Lazuli | Level 10

## Re: How can i obtain a relative risk for my multi variable linear regression analysis of my bio-medi

I am afraid that you have not understood the concept of "relative risk" that clearly. When we talk about relative risks, we usually divide the entire study population into distinct groups by a certain factor and then compare the incidence of an event by dividing them. In other words, the outcome, i.e., the dependent variable, should be categorical (usually binary). But you said the dependent variable in your model is continuous. There is no way of calculating relative risks when that is the case.

Discussion stats
• 2 replies
• 410 views
• 0 likes
• 3 in conversation